extraterrestrial organic matter preserved in 3.33â€¯ga

Available online at www.sciencedirect.com

www.elsevier.com/locate/gca

ScienceDirect

Geochimica et Cosmochimica Acta 258 (2019) 207–225

Extraterrestrial organic matter preserved in 3.33 Gasediments from Barberton, South Africa

Didier Gourier a,⇑, Laurent Binet a, Thomas Calligaro f,a, Serena Cappelli a

Herve Vezin b, Jean Breheret c, Keyron Hickman-Lewis d,g, Pascale Gautret e

Frederic Foucher d, Kathy Campbell d, Frances Westall d

aChimie ParisTech, PSL University, CNRS, Institut de Recherche de Chimie de Paris (IRCP), F-75005 Paris, FrancebUniversity Lille, CNRS, UMR 8516 – LASIR - Laboratoire de Spectrochimie Infrarouge et Raman, F-59000 Lille, FrancecGeoHydrosystemes Continentaux (GeHCO), EA 6293, Universite de Tours, Parc de Grandmont, F-37200 Tours, France

dCentre de Biophysique Moleculaire (CBM), CNRS, F-45071 Orleans, FranceeUniversite d’Orleans BRGM, CNRS, Institut des Sciences de la Terre d’Orleans (ISTO), 45060 Orleans, France

fCentre de Recherche et de Restauration des Musees de France (C2RMF), Palais du Louvre, F-75001 Paris, FrancegDipartimento di Scienze Biologiche, Geologiche e Ambientali, Universita di Bologna, Via Zamboni 67, I-40126 Bologna, Italy

Received 20 September 2018; accepted in revised form 6 May 2019; Available online 15 May 2019

Abstract

Electron paramagnetic resonance (EPR) analysis of carbonaceous, volcanic, tidal sediments from the 3.33 Ga-old JosefsdalChert (Kromberg Formation, Barberton Greenstone Belt), documents the presence of two types of insoluble organic matter(IOM): (1) IOM similar to that previously found in Archean cherts from numerous other sedimentary rocks in the world andof purported biogenic origin; (2) anomalous IOM localized in a 2 mm-thick sedimentary horizon. Detailed analysis bycontinuous-wave-EPR and pulse-EPR reveals that IOM in this layer is similar to the insoluble component of the hydro-genated organic matter in carbonaceous chondrites, suggesting that this narrow sedimentary horizon has preserved organicmatter of extraterrestrial origin. This conclusion is supported by the presence in this thin layer of another anomalous EPRsignal at g = 3 attributed to Ni-Cr-Al ferrite spinel nanoparticles, which are known to form during atmospheric entry of cos-mic objects. From this EPR analysis, it was deduced that the anomalous sedimentary layer originates from deposition, in anearshore environment, of a cloud of tiny dust particles originating from a flux of micrometeorites falling through the oxygen-poor Archean atmosphere.� 2019 Elsevier Ltd. All rights reserved.

Keywords: Extraterrestrial organic matter; Spinels; Electron paramagnetic resonance; Early Archean; Josefsdal chert formation

1. INTRODUCTION

It is widely believed that most of the water and organicmatter on the Hadean Eon (4.54–4 Ga) was of extraterres-

https://doi.org/10.1016/j.gca.2019.05.009

0016-7037/� 2019 Elsevier Ltd. All rights reserved.

⇑ Corresponding author.E-mail address: [email protected]

(D. Gourier).

trial material (Morbidelli et al., 2012; Marty et al., 2013 andreferences therein). While recent models suggest that themajority of extraterrestrial flux occurred before 4.4 Ga(Boehnke and Harrison, 2016; Genda et al., 2017), thereis nonetheless evidence for continued input of extraterres-trial material throughout the Early Archaean (�4–3.2 Ga). For example, Schoenberg et al. (2002) documentevidence for meteoritic input through tungsten isotope


mailto:[email protected]


http://crossmark.crossref.org/dialog/?doi=10.1016/j.gca.2019.05.009&domain=pdf

208 D. Gourier et al. /Geochimica et Cosmochimica Acta 258 (2019) 207–225

analysis of 3.7 Ga metamorphosed sediments from Isua,Greenland, while a number of layers of impact spheruleshave been observed in Early-Mid Archean (�3.5–3.2 Ga)sediments from the Barberton Greenstone Belt (SouthAfrica) and the Pilbara Craton (Western Australia)(Glikson et al., 2004; Gomes et al., 2005; Krull-Davatzeset al., 2010; Bottke et al., 2012; Lowe et al., 2014). This fluxstill continues today, albeit to a lesser extent (e.g., Love andBrownlee, 1991; Yada et al., 2004).

Extraterrestrial material may have provided substantialsources of complex organic molecules for the emergenceof life (e.g. Cooper et al., 2001; Strasdeit, 2005; Pasek andLauretta, 2008). Organic matter in carbonaceous chon-drites can reach up to 4% (Sephton, 2002) and up to 85%in ultracarbonaceous micrometeorites (Duprat et al.,2010; Dartois et al., 2018). About 75% of organic matterin carbonaceous chondrites is refractory, consisting mostlyof small aromatic moieties linked by short and branchedaliphatic chains (Derenne and Robert, 2010). This insolublecomponent will hereafter be referred to as insoluble organicmatter (IOM). Although representing only a small fraction,soluble components exhibit a vast range of composition,with more than 14,000 different molecules containing C,H, O, N and S having been identified to date (Schmitt-Kopplin et al., 2010; Remusat, 2014). Meteoritic debrisand micrometeorites falling into the ocean of the earlyEarth would have been rapidly broken up and partially dis-solved in the globally acidic and warm ocean (Pinti, 2005;Tartese et al., 2017), thus liberating both the soluble andthe refractory fractions (Westall et al., 2018). The solublemolecules would be rapidly dissolved into seawater whereasthe refractory materials (the IOM) would remain in thedetrital fraction of the volcanic sediments coating the floorof the ocean and the numerous platform-covering (sub-merged continents) shallow seas. Of particular interest isthe fate of this latter fraction, its potential involvement inprebiotic reactions leading to the emergence of life, andits distinction from organic biosignatures of ancient tracesof life. In this study, we describe evidence of the oldesttraces of extraterrestrial organic matter yet identifiedthrough a detailed, sedimentological, geochemical, andElectron Paramagnetic Resonance (EPR) study of EarlyArchaean (3.33 Ga) sediments in the Barberton GreenstoneBelt (BGB). These results will certainly have implicationsfor the search for organic biosignatures of primitive lifeon Earth and on Mars

2. THE JOSEFSDAL CHERT AND ITS

SEDIMENTOLOGICAL CONTEXT

The 3.33 Ga Josefsdal Chert (the stratigraphic equiva-lent of unit K3c in the Kromberg Formation), located inthe southern part of the Barberton Greenstone Belt(Fig. 1A) (Lowe et al., 2012) is an extensive deposit ofmostly basaltic volcanic sediments deposited on top ofhydrothermally silicified basalts extruded onto a shallowwater platform (Westall et al., 2006, 2011, 2015). The 7 mto up to 20 m thick deposit (variations in thickness depend-ing upon underlying, fault-bounded topography) is dividedinto four stratigraphical units denoted 1–4 comprising four

recurring lithological facies denoted A, B, C and D (Westallet al., 2015) (Fig. 1B). The sedimentological structures andtextures indicate deposition in an initially regressive envi-ronment from upper offshore to upper shore face (suprati-dal), followed by a transgressive sequence back toshoreface/foreshore. The sediments are highly silicified(85.5–99.9%), the silicification occurring more or less con-temporaneously with deposition. The clastic sedimentscomprise volcanic protoliths including lithic fragments ofbasalt, volcanic glass, accretionary lapilli, minerals suchas feldspar laths, pyroxene remnants, rare spherulites, aswell as carbon of both demonstrably biogenic and detritalorigins. The lithic fragments are generally highly alteredto muscovite and anatase and replaced by silica, althoughsome remnant structures such as tear-drop shapes, remain.Chemical deposits in the form of silica gels are both inter-calated with the detrital sediments as a cement and formlayers precipitated on top of bedding planes.

Biogenic features have been documented in variousdegrees of abundance throughout the sedimentary succes-sion (Westall et al., 2006, 2011, 2015). They comprise thecarbonaceous and silicified remains of delicate pho-totrophic biofilms and mats formed atop sediment layers,as well as densely clotted mats of thickly colonized volcanicparticles. Biogenicity interpretations were made on thebasis of morphology, carbonaceous composition, carbonisotopic signature consistent with microbial fractionation,molecular composition (as determined by ToF-SIMS), allwithin a specific context defined on the mineralogic to localgeological environment. Further details regarding method-ology and discussion of biogenicity criteria are detailed inthe above published papers.

We here concentrate on Facies D of Unit 3 in the Josefs-dal Chert (Fig. 1), and use a black chert from Facies C onlyfor comparison. Unit 3, from which the studied samples aresourced, is bounded below and above by conspicuous layersof intraformational breccia. Facies D comprises finely lam-inated green/grey and black sediments (ranging from mm’sto cm’s in thickness) consisting of repetitive bundles of vol-canic ash and fine accretionary lapilli deposited as parallel,graded beds, the tops of which are delineated by very finecarbonaceous laminae. The carbonaceous component rep-resents deposition in relatively quiet periods between ashfallepisodes (Supplementary Fig. A1). Intermittent remobiliza-tion of the sediments by current and wave action resulted incross bedding and oscillatory bedding, the lamina bundlesbeing bounded between two of such deposits. These sedi-mentary structures, as well as mud cracks (not shown),are suggestive of an upper shore face (subtidal-tidal) envi-ronment in which the sediments were occasionally subjectedto subaerial exposure. The fact that it is possible to traceindividual layers, even the thinnest laminae (mm-thickness), over distances of several kilometers (Westallet al., 2015) suggests that deposition was of regional scalefrom ash fall with only occasional current reworking, i.e.during quiescent hydrodynamic regimes. In addition tothe underlying and overlying breccia layers, Facies D exhi-bits evidence of more significant disturbance in the form ofslumping and a certain amount of blurring of the sedimen-tary structures indicating liquefaction. These structures

Fig. 1. Location and stratigraphy of the 3.33 Ga Josefsdal Chert in the Barberton Greenstone Belt (BGB), South Africa. (A) Geographicaland geological location of the BGB and (B) location of the studied area within the southern part of the BGB (after Westall et al. (2015)). (C)Composite field photographs of Facies D, occurring in Unit 3 (Fig. 1D), in which the carbonaceous layer F of interest to this study is markedby a white arrow. The sample analysed and shown in Fig. 4 is located by a rectangular box in this image. Facies D comprises a series of finelylaminated, graded tuff layers, bounded by wave/current reworked intervals. Carbonaceous layer F occurs in the 7th bundle of tuff layers(group marked as 7). (D) Generalized stratigraphic column with sedimentary units noted 1–4 with their environmental assignations. The faciesassociations A–D are noted by different colors.

D. Gourier et al. /Geochimica et Cosmochimica Acta 258 (2019) 207–225 209

were caused by shocks, possibly seismic and/or explosivehydrothermal events.

Of the eight bundles of laminae comprising the Facies Dstrata of Unit 3, the horizon labelled F in Fig. 1D,occurring towards the base of the 7th bundle, is the mainfocus of our attention in this study (SupplementaryFig. A2). Horizon F is a 2 mm thick composite carbona-ceous layer that formed at the top of a graded tuffaceous

bed. It comprises three separate, carbon-rich laminae(labelled F0, F00, and F00 0 in Fig. 2). As shown in this paper,lamina F0 is the one in which the EPR peak indicatingextraterrestrial organic matter is strongest. This 200–300 mm thick lamina is a mixture of carbonaceous matter,highly altered and silicified spherical clasts (accretionarylapilli and both barred and structureless spherulites), andincludes a small component (<0.2%) of minute pyrite

Fig. 2. Sedimentological detail of horizon F. (A) Outcrop detail showing that horizon F occurs between two layers of graded tuffaceoussediments. It consists of three laminae, labelled F0, F00, F00 0. (B) Optical microscope view of a thin section (60 mm) of horizon F showing thatlaminae F0 and F00 0 comprise spherulite-rich layers with intermixed carbonaceous matter, as well as fine, undulating carbonaceous laminaebelow and above. Lamina F00, on the other hand, is thinner and laterally discontinuous. The red box denotes the area analyzed by Ramanspectroscopy (D, E). The white arrow marks the location of a prominent, altered spherule visible in (B–E). (B0) The inset shows spherules inlamina F0 having a barred texture, similar to extraterrestrial spherules described by Simonson and Glass (2004). (C) Optical detail of areamapped by Raman. (D) Combined composition of the mapped area. Carbon is in green, quartz in yellow/orange, muscovite in pink, anatasein blue. (E) The three carbonaceous laminae are highlighted in different colors (F0 in green, F00 in blue and F00 0 in red). (F) Raman spectra of therespective three laminae F0, F00 and F00 0 show that all spectral parameters are identical: D and G peaks are located at 1354 cm�1 and1598 cm�1; their full width at half maximum values are equal to 50 cm�1 and 57 cm�1 respectively, and the D/G intensity ratio is equal to 2.2.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)


crystallites. It is separated from the second carbonaceouslamina (F00) by about 400 mm of tuffaceous sediments. Lam-ina F00 is very fine (�50 mm), undulating and laterally dis-continuous (Fig. 2B). A 600 lm tuffaceous layer separateslamina F00 from the third carbonaceous lamina F00 0 which,like F0, is a mixture of altered spherulites and carbonaceousmatter (Fig. 2B).

3. SAMPLES AND METHODS

3.1. Samples

Samples were obtained during numerous field cam-paigns. Dedicated field work to elucidate the geologicalcontext was undertaken in 2012 and 2014. Two samples,collected in upper shoreface facies C and D of unit 3(Fig. 1) were selected for study by EPR. The first sample,labelled a (size 5.7 cm � 2 cm) was collected from the tuffa-

ceous layer of Facies D marked by an arrow in Fig. 1C.This sample, shown in Figs. 4 and 6, contains four of theeight black horizons of the 7th bundle, labelled E, F, Gand H, horizon F being the carbonaceous horizon of inter-est in this study. It was cut for EPR analysis into 40 slices(1–1.5 mm width; numbered from 0 to 39) parallel to thesedimentation plane. The second sample (�3 cm � 2 cm),labelled b, was collected from the homogenous and rela-tively carbon-rich Facies C, which underlies Facies D(Fig. 1D). It was used as a reference for the EPR signatureof carbonaceous matter in the Josefsdal Chert in general. Itwas polished and cut into 27 strips (section 2 � 2 mm2, 3–8 mm length) for EPR analysis (see Fig. 6).

3.2. Methods

Thin sections of the same samples were made for opticalmicroscopy and Raman analyses. Optical microscopy was


conducted on 30–60 mm polished thin sections using anOlympus BX51 microscope at the CNRS-CBM (Orleans).Raman analyses were undertaken at CNRS-CBM (Orleans)on the polished thin sections using a WITec Alpha500 RARaman spectrometer equipped with green laser (frequency-doubled Nd:YAG laser at 532 nm) and a Nikon E Plan 50xobjective of numerical aperture 0.75. The analysis was con-ducted with the method described by Foucher et al. (2017),with a laser power chosen to have an energy density lowerthan 1 mJ/mm3 to prevent sample heating. Raman mappingwas carried out on a sample containing horizon F collectedclose to the sample a location, but not sample a itself. Thereason for this separate analysis was to avoid perturbationsof the organic matter of the black horizon F by the Ramanlaser beam, which would invariably modify the EPR spec-trum of IOM. The Raman analysis documents the perva-sive alteration of the volcanic protoliths to muscovite andanatase/rutile followed by significant silicification (>85–96% SiO2) (Westall et al., 2018). Carbonaceous matterwas concentrated in the horizon F laminae (Fig. 2C–E).Raman spectra of minerals in horizon are given in Supple-mentary Fig. B1. In detail, Raman spectroscopy also showsthat the carbonaceous component is finely disseminatedbetween the volcanic clasts in lamina F0 and F00 0. The verythin lamina F00 comprised primarily carbonaceous matter(IOM) in a quartz matrix. Lamina F00 is absent in samplea. Parameters of the Raman spectra show that there areno differences in peak position, shape, width and ID/IGratio of D and G Raman bands of the IOM in the threehorizons (Fig. 2F).

Elemental micromapping of sample a was carried outusing Particle Induced X-ray Emission (PIXE) ion micro-probe analysis (Ryan, 1995) at the Accelerator Grand Lou-

vre for Elemental Analysis (AGLAE) of the Centre de

Recherche et de Restauration des Musees de France

(C2RMF) located in the Palais du Louvre, Paris(Calligaro et al., 2011). A rectangular area of 5 � 0.5 mm2

from horizon F was scanned at a resolution of 20 lm usinga 3-MeV proton beam of �30 lm diameter extracted in theair through a 100-nm thick Si3N4 membrane, yielding mapsof 250 � 25 points. Three 50-mm2 SDD detectors (KetekAXAS-A) were placed 20 mm from the sample at 50� rela-tive to the beam axis and screened with 100-mm Al absorberfoils to collect trace element X-rays (Z > 20). An additionaldetector without absorber fitted with a magnetic deflectorin a helium gas flow at 2 litre/min was used to record bulkelement X-rays, mainly Si, S and Fe required for the traceelement calibration. Each PIXE spectrum was collectedwith a proton dose of 32 nC. The mean beam intensitywas 16 nA and the total mapping time was 3.5 h. Countrates in major elements and trace element detectors were20 and 1.5 kcounts/sec, respectively. Hence pile-up was neg-ligible. The acquired spectra were processed to obtain quan-titative composition using the TRAUPIX program (Pichonet al., 2014) built upon the GUPIX software package(Campbell et al., 2010). Processing parameters wereadjusted using the Diorite DR-N reference geostandard(Govindaraju, 1982). It is important to note that this PIXEanalysis was performed on the remaining part of sample aafter the collection of the 40 slices for EPR analyses. This

precaution was necessary to avoid any change to theIOM by the ion beam, which would invariably perturbthe EPR signatures.

Continuous wave EPR spectra (cw-EPR) were recordedat the X-band (�9.4 GHz) and at room temperature using aBruker ELEXSYS E500 spectrometer equipped with a4122SHQE/011 resonator. The 67 fragments of cherts wereseparately introduced into EPR quartz tubes for cw-EPRanalysis. Pulsed-EPR experiments were carried out at 5 Kwith a Bruker ELEXSYS E500 X-band spectrometerequipped with a Bruker cryostat ‘‘cryofree” system. Hyper-fine Sublevel Correlation pulse sequence (HYSCORE) wasused to reveal interactions of the electron spins of carbona-ceous matter with 13C (I = 1/2; 1.1 % abundance), 1H(I = 1/2; 100% abundance), and 31P (I = 1/2; 100% abun-dance) nuclei. Using this technique, a spin echo is generatedby the pulse sequence p/2-s-p/2-t1-p-t2-p/2-s-echo. Theangles p/2 and p represent the flip angles of the electronmagnetization. Its intensity is measured by varying thetimes t1 and t2 at constant time s in a stepwise manner.The lengths of the p/2 and p pulses were fixed at 16 nsand 32 ns, respectively. 256 � 256 data points were col-lected for both t1 and t2 at increments of 20 ns. s valuewas set at 136 ns for all samples. The unmodulated partof the echo was removed by using second-order polynomialbackground subtraction. The magnitude spectrum wasobtained after 2D Fourier transformation of the spectraby using a Hamming apodization function. Pulse-EPRanalysis was performed on slice 25 of sample a and frag-ment 9 of the carbon-rich sample b (see Fig. 6). For samplea, due to weak echo signal even at 5 K, the spectrum con-sists in 100 accumulations with 100 shots per points. Orig-inal EPR and PIXE data are given in Electronic annexes.

4. RESULTS

4.1. General EPR features of Josefsdal chert

The EPR spectra of the 40 slices of the laminated chert(sample a) vary significantly according to their stratigraphic(vertical) position. Fig. 3 shows two examples of EPR spec-tra corresponding to slice 33 (located above black horizonE) and slice 25 which contains lamina F0 of horizon F.The locations of these slices in the sample are shown inFig. 4. Slice 33 (Fig. 3A and B) is lithologically representa-tive of the bulk sample, except for the relative intensity ofthe EPR signals, which varies from slice to slice. The mostprominent feature in Fig. 3A is a broad symmetrical lineat g � 2 which broadens and distorts at low temperature(not shown). This peculiar temperature dependence indi-cates that this signal corresponds to the resonance signalof superparamagnetic (SPM) nanoparticles, possibly fer-rites. Such particles become ferrimagnetic at low tempera-ture, which is responsible for the broadening, shifting anddistortion of the line (Griscom, 1980, 1984). At low mag-netic field, an EPR line at g = 4.3 is typical of Fe3+ impuri-ties in low symmetry sites of inorganic hosts. Zooming in thehigh field portion of the SPM line reveals four additionalEPR signals (Fig. 3B). The weak narrow line at g = 2.003,noted as IOM on the diagram, is due to carbonaceous

Fig. 3. cw-EPR spectra of two representative fragments of sample a: slice 33 (A and B) and slice 25 (C and D). (A) Full range EPR spectrumof slice 33 showing Fe3+ impurities and a strong signal of superparamagnetic resonance (SPM) at g � 2 due to the presence of magneticnanoparticles. The rectangular box represents the part of the EPR spectrum detailed in (B), which shows the weak line of carbonaceous matter(IOM), the E0 center and the Al3+-O� center. The spectrum of the latter recorded at 50 K is shown in the inset, where the hyperfine interactionwith the 27Al nucleus is clearly seen. (C) Total EPR spectrum of slice 25, corresponding to lamina F0 of horizon F. The rectangular boxrepresents the part of the EPR spectrum detailed in (D), which shows the IOM line. The two EPR signals that present extraterrestrialcharacteristics (the IOM line and the SPM signal at g = 3) are represented in red. Experimental conditions: microwave power 10 mW,modulation depth 0.1 mT, microwave frequencies 9.7443 GHz (spectra A and B), 9.7418 GHz (spectra C and D); inset of panel B: microwavepower 2 mW, modulation depth 0.1 mT, microwave frequency 9.3924 GHz. (For interpretation of the references to colour in this figurelegend, the reader is referred to the web version of this article.)


matter (Skrzypczak-Bonduelle et al., 2008). A narrow peakin the high field flank of this IOM line is due to oxygenvacancy centers (E0 centers) in the SiO2 matrix (see Weil,1984; Weeks et al., 2008 and references therein). This attri-bution to E0 centers was demonstrated by the fact that theE0 line can be detected selectively by recording the EPRspectrum 90� out-of-phase with the modulation field (notshown), which is not the case for other EPR signals(Skrzypczak-Bonduelle et al., 2008). A broad and unre-solved signal is also visible in the low field side of the IOMline, which is best resolved by recording the EPR spectrumat 50 K (insert in Fig. 3B). This signal exhibits a character-istic structure due to hyperfine interaction with 27Al nuclei(nuclear spin I = 5/2) and three components, gz = 2.0060gx = 2.0085 and gy = 2.003, of the g-factor. This was alsoobserved in carbonaceous cherts of the WarrawoonaGroup, Pilbara Craton, Australia (Skrzypczak-Bonduelleet al., 2008). This paramagnetic center is an (O�-Al3+)center, resulting from hole trapping by an oxygen ion

(i.e a O� center) adjacent to an Al3+ impurity in Si site ofa silicate matrix (Weil, 1984, Botis and Pan, 2009). Note thatthe aluminum is most probably related to the presence ofmuscovite, the lower greenschist-metamorphic variant ofsmectite, a common, aqueous alteration product of volcanicdetrital precursors that was subsequently silicified (Westallet al., 2015). In the high magnetic field part of the spectrum,a doublet of EPR lines is due to an as yet unidentified center(referred to as the X center) in the mineral matrix, and willnot be discussed further in this paper as it is ubiquitous inthe sample.

The EPR spectrum of slice 25 (Fig. 3C), correspondingmostly to lamina F0 of horizon F, exhibits unusual featurescompared to other slices of sample a. First, the SPM signalof superparamagnetic nanoparticles (g = 2) is lacking andthe IOM line of carbonaceous matter is more intense. Sec-ond, in addition to the signal at g = 4.3 of Fe3+ impurities,present throughout in the sample, there is another signal atg = 3, which is of particular interest because it was clearly


observed only in horizon F of sample a (slices 23–26). Thecentral part of the spectrum (blue box in Fig. 3C) clearlyshows the IOM line as well as the undetermined X center(Fig. 3D).

The heterogeneity of sample a is reflected in the strongvariability of the EPR spectra throughout the 40 slices.Fig. 4 presents the weight normalized EPR intensity profileof four EPR signals: the SPM signals at g = 2 and g = 3(Fig. 4A), the signal of Fe3+ impurities at g = 4.3(Fig. 4B) and the IOM line (Fig. 4C). The peak-to-peaklinewidth of the IOM line is shown in Fig. 4D. The slicenumber is given in abscissa. EPR intensities of each typeof EPR signal are given in arbitrary units, which do notreflect their abundance. For example, Fe3+ impurities inmineral hosts (signal g = 4.3) are present everywhere inthe sample, but their average concentration is low (proba-bly no more than 100 ppm) because the total Fe contentin the bulk of sample a (except F horizon) is less than1000 ppm. On the contrary, SPM particles are inhomoge-neously distributed. The SPM signal at g = 2 is present onlyin slices 0–12 and slices 30–36, corresponding to the blackhorizons G-H and E, respectively (Fig. 4A). Interestingly,the signal at g = 3 (in red in Fig. 4A) is observed only inslices 23–26, corresponding to the black horizon F, whichalso contains an anomalous IOM line. This will provide afocus for discussion in the following sections.

4.2. EPR lineshape analysis of carbonaceous matter

EPR associated with carbonaceous matter (IOM) havelong been studied, for example in ancient carbonaceousmaterial such as coal (Uebersfeld and Erb, 1956;Retcofsky et al., 1968; Mrozowski, 1988a,b), carbonaceouscherts (Skrzypczak-Bonduelle et al., 2008; Bourbin et al.,2013), and carbonaceous chondrites (Duchesne et al.,1964; Vinogradov et al., 1964; Schultz and Elofson, 1965;

3

Fig. 4. Variation of cw-EPR spectra in the 40 slices of sample ashown at the bottom of the diagram (the location of this sample inthe context of the outcrop is shown by a rectangle in Fig. 1C). Theslices, labelled 0–39 along the abscissa of (D), were collectedparallel to the stratification. In terms of stratigraphy, slice 0 (to theleft) corresponds to the stratigraphic top of the sample and slice 39to the bottom (to the right). The image of sample a is shown withthe position of the 40 slices studied by EPR (vertical bars) and fourblack horizons (labelled E to H). The white box represents the areascanned in the PIXE analysis (see Section 4.4.1). Data for the blackhorizon F (slices 23–26) are highlighted with a blue rectanglecovering panels A–D. Slice 25 comprises a large part of lamina F0

of horizon F and is highlighted by a discontinuous vertical line. (A)EPR intensity profile of the superparamagnetic resonance signal(SPM) of possibly ferrite nanoparticles (g = 2), and of the signal atg = 3 attributed to Ni-Cr-Al ferrite nanoparticles; (B) EPRintensity profile of Fe3+ impurities (g = 4.3), (C) EPR intensityprofile of the IOM line; (D) Variation of the peak-to-peaklinewidth of the IOM line. All intensities are given in arbitraryunits. (For interpretation of the references to colour in this figurelegend, the reader is referred to the web version of this article.)

Fig. 5. EPR line of carbonaceous matter (IOM) in slice 17 ofsample b (A) and slice 25 (mostly lamina F0) of sample a (B). Theexperimental spectrum (black line) is compared with calculatedspectra (dashed lines) corresponding to Lorentzian (R10 = 0, red)and stretched Lorentzian profiles with R10 = �2.42 (blue) andR10 = �3 (green). (For interpretation of the references to colour inthis figure legend, the reader is referred to the web version of thisarticle.)


Binet et al., 2002, 2004; Gourier et al., 2008; Delpoux et al.,2011). These carbonaceous materials invariably give a sin-gle EPR line due to aromatic radical moieties with anunpaired electron spin delocalized in p-type molecular orbi-tals (Uebersfeld et al., 1954; Retcofsky et al., 1968;Mrozowski, 1988b; Dickneider et al., 1997). The IOM linecan be characterized by its intensity (proportional to thenumber of radicals), its g-factor (measured from the mag-netic field at resonance) and its lineshape.

Compared to the homogenous, carbon-rich black chert(sample b, Figs. 5A and 6) that is used as a reference andwhich exhibits a sharp IOM line, the laminated chert (sam-ple a) has only a weak IOM line, even in the black horizonsG, H and E. The IOM line is relatively strong only in slice25, which mainly corresponds to lamina F0 of black horizonF (Figs. 4C and 5B). The peak-to-peak linewidth DBpp ofthe IOM line also varies across the sample, as shown inFig. 4D. Three different linewidth domains can be distin-guished. The IOM lines of the majority of sample a arecharacterized by DBpp in the range of 0.25–0.4 mT, whichagrees with values of 0.2–0.3 mT previously measured in�3.4 Ga-old cherts from the BGB (Skrzypczak-Bonduelleet al., 2008). The IOM line in slices 3–7 exhibits a very

narrow linewidth DBpp = 0.13 mT, previously observed ina 3.48 Ga-old chert from the Dresser Formation (Warra-woona Group, Australia) (Skrzypczak-Bonduelle et al.,2008). In contrast, the IOM line of slice 25, dominated bylamina F0, has a linewidth DBpp = 0.5 ± 0.02 mT,which exceeds the largest linewidth previously measuredin Archean cherts but corresponds to the valuesDBpp = 0.38–0.54 measured for the IOM line in carbona-ceous chondrites (Binet et al., 2002, 2004). The broadeningof the IOM line can be ascribed to two different origins. ForPhanerozoic, Proterozoic and meteoritic IOM, the line-width is determined by the hydrogen/carbon ratio, whilefor Archean carbonaceous matter, the linewidth is mainlydetermined by the metamorphic grade (Binet et al., 2002;Skrzypczak-Bonduelle et al., 2008).

Additional information on the origin of this IOM can beobtained from analysis of the IOM lineshape. We have pre-viously observed that the shape of IOM lines in carbona-ceous cherts varies from Gaussian-Lorentzian (Voigtshape) to Lorentzian for ages ranging from Phanerozoicto Proterozoic, then to stretched Lorentzian for ages datingback to Archean. Skrzypczak-Bonduelle et al. (2008) pro-posed a numerical analysis of the IOM line by deriving alineshape parameter (noted R10) defined to quantify thedeviation of the signal from a Lorentzian shape (seeSkrzypczak-Bonduelle et al., 2008; Bourbin et al., 2013for more detail). Briefly, the lineshape parameter is charac-terized by values R10 = 0 for a pure Lorentzian line, R10 > 0for an increasing Gaussian character, and �3 < R10 < 0 fora stretched Lorentzian line. Fig. 5 shows the analysis of twoIOM lines, corresponding to slice 25 of the laminated cherta (lamina F0 of black horizon F), and to a representativefragment of the carbon-rich chert b. Both experimentalspectra are compared with a Lorentzian line (R10 = 0)and two stretched Lorentzian lines (R10 = �2.42 and �3).The IOM line of chert a (slice 25) is close to Lorentzian(Fig. 5B), while the Lorentzian is clearly stretched in chertb, with R10 = �2.42 (Fig. 5A).

From lineshape parameters R10 and linewidths DBpp col-lected from previous EPR analyses of Precambrian cherts(Skrzypczak-Bonduelle et al., 2008; Bourbin et al., 2013)and carbonaceous chondrites (Binet et al., 2002, 2004), weconstructed an EPR diagram of primitive carbonaceousmatter, plotting the lineshape parameter R10 versus the line-width parameter DBpp (Fig. 6). Open circles labelled by let-ters a to l represent samples collected from differentgeological settings and chondrites, detailed in the captionof Fig. 6. Archean, Proterozoic and meteoritic IOMs aredistributed in three different domains in this diagram, high-lighted by colored rectangles. Archean IOM (blue area) ischaracterized by a stretched Lorentzian shape(�2 � R10 � �3), while Proterozoic IOM (orange area) ischaracterized by a shape close to the Lorentzian(�1 � R10 � +1). Meteoritic IOM (green area) andProterozoic IOM have similar lineshapes but differentlinewidths, that of meteorites being higher. It is importantto note that the EPR of most Phanerozoic IOM (youngerthan 450 Ma) fall out of the field of this phase diagram,as demonstrated by the 45 Ma Clarno sample (Fig. 6, datapoint i), because the EPR signal of organic matter in such

Fig. 6. EPR lineshape factor R10 versus peak-to-peak linewidthDBpp of IOM lines for carbonaceous chondrites and Precambriancherts (open circles, data from Skrzypczak-Bonduelle et al., 2008;Bourbin et al., 2013; Binet et al., 2002, 2004), and for IOM line insample b (green circle) and sample a (red circles, slices 25 and 33) ofJosefsdal Chert. The diagram is constructed from data obtained forthe following samples: (a) Dahongyu Formation, Hebei, China(1.43–1.77 Ga); (b) Gunflint Formation, Shreiber Beach locality,Ontario, Canada (1.88 Ga); (c) Duck Creek, Wyloo Group,Ashburton Trough, Australia (1.85–2.2 Ga); (d) Jeerinah Forma-tion, Fortescue Group, Hamersley basin, Australia (2.75 Ga); (e)Middle Marker, Komati Formation, Barberton Greenstone Belt,South Africa (3.47 Ga); (f) Dresser Formation, WarrawoonaGroup, Pilbara craton, Australia (3.48 Ga); (g) Kromberg Forma-tion, Upper Onverwacht Group, Barberton Greenstone Belt, SouthAfrica (3.42–3.45 Ga); (h) Middle Marker, Komati Formation,Barberton Greenstone Belts, South Africa (3.47 Ga); (i) ClarnoFormation, John Day Basin Tectonic Unit, Oregon, USA(0.045 Ga); (j) Orgueil meteorite; (k) Paris meteorite; (l) Murchisonmeteorite; (m) Tagish Lake meteorite. The two images representsample a (with black horizons E, F, G H), and sample b polishedand cut in 27 slices (slice 9 is missing). The error bars for sample brepresent the dispersion of the values measured for the 27 slices.(For interpretation of the references to colour in this figure legend,the reader is referred to the web version of this article.)


samples often exhibits a resolved hyperfine interaction withhydrogen nuclei, giving a multi-line EPR pattern(Skrzypczak-Bonduelle et al., 2008).

Data from the Josefsdal cherts in Fig. 6 suggest thatsome of the IOM of samples b (facies C) and a (facies D)has different origins. First, the values measured for the 27slices of the carbon-rich sample b are almost identical andlocalized in the Archean part of the diagram (green circlein Fig. 6), the error bars representing the dispersion ofexperimental values. This result indicates that the structureand evolution of this IOM is similar to that of otherArchean IOM in carbonaceous cherts from South Africaand Australia. On the contrary, data from the carbon-poor sample a are more widely scattered. Red circles inFig. 6 represent experimental values for slices 25 and 33of sample a. Although the IOM in slice 33 is at the outeredge of the domain covered by other Archean samples, itslineshape evidently corresponds to the stretched Lorentziantypical of Archean IOM. Surprisingly, the data for slice 25,consisting mainly of IOM from lamina F0, is located clearly

outside the domain of Archean IOM and inside the field ofcarbonaceous chondrites.

This feature is a first indication of a possible extraterres-trial origin for the IOM in horizon F.

4.3. Pulse-EPR analysis of carbonaceous matter

The proposed extraterrestrial origin of the IOM in laminaF0 is supported by pulse-EPR analysis. UsingHyperfine Sub-level Correlation (HYSCORE) spectroscopy (Schweiger andJeschke, 2001) we identified nuclei with non-zero nuclearspin (mainly 1H, 13C, 31P) interacting with the electron spinsof IOM radicals. Previous studies showed that meteoriticIOM and biogenic IOM in cherts exhibit differentHYSCORE spectra resulting from different molecular struc-tures and hydrogen contents, which in turn reflect their ori-gin and diagenetic/metamorphic evolution (Gourier et al.,2008, 2010, 2013). Fig. 7 shows HYSCORE spectra of slice25 of sample a (Fig. 7E) and a fragment of sample b(Fig. 7F) of the Josefsdal Chert, compared with IOM ofthe Orgueil meteorite (Fig. 7D), two carbonaceous chertswith indisputably biogenic IOM: Clarno Formation, Ore-gon, USA (Eocene, 0.045 Ga, Fig. 7A), Gunflint Formation,Ontario, Canada (Paleoproterozoic, 1.88 Ga, Fig. 7B), andone chert from the Dresser Formation of the WarrawoonaGroup, Pilbara, Australia (Archean, 3.48 Ga, Fig. 7C)(Gourier et al., 2008, 2013), with IOM of likely biogenic ori-gin. HYSCORE spectra of Clarno and Gunflint cherts,which contain well-preserved fossils, show similar character-istics, namely (i) a non-linear hydrogen ridge centred at thenuclear frequency 14.5 MHz of 1H, (ii) a very weak signalat the nuclear frequency 3.7 MHz of 13C and (iii) a double-frequency signal at 12 MHz, which is twice the nuclearfrequency 6 MHz of 31P. Phosphorus is one of the bio-essential elements, and phosphate are also associated tofossil bacteria via biomineralization (Li et al., 2013;Cosmidis et al., 2015). So it is not surprising to detect 31Presonance associated to carbonaceous fossils. HYSCOREspectra of meteoritic IOM from Orgueil differ from thoseof biogenic IOMby the presence of a linear 1H ridge, a broadand intense 13C ridge, and the lack of a double-frequency 31Psignal. The same features were observed for the Tagish Lakemeteorite (Gourier et al., 2008, 2013). The differencesobserved for 1H and 13C signals are most probably relatedto the molecular structure of the IOM.Meteoritic IOM com-prises small polyaromatic units linked by short and highlybranched aliphatic moieties (Derenne and Robert, 2010),while biogenic IOM contains polyaromatic units with fewerbranched aliphatic side chains (Gourier et al., 2013). TheHYSCORE spectrum of the Archean chert of the DresserFormation is clearly different from that of more recent cherts(Clarno and Gunflint Formations) due to the absence of the1H signal. This is because the hydrogen/carbon ratiodecreases during the maturation of the IOM, approachingzero in mature Archean IOM. Thus, only the double-frequency signal of 31P nuclei is present, suggesting thatthe IOM is derived from bacterial sources. The same spectro-scopic characteristics observed for the 3.48 Ga DresserFormation chert (Fig. 7C) are also clearly present in faciesC (sample b) of the Josefsdal Chert (Fig. 7F), which shows

Fig. 7. HYSCORE spectra at 5 K of (A) a chert from the Clarno Formation, John Day Basin Tectonic Unit, Oregon, USA (0.045 Ga); (B) achert from the Gunflint Formation, Schreiber Beach locality, Ontario, Canada (1.88 Ga); (C) a chert from the Dresser Formation,Warrawoona Group, Pilbara craton, Australia (3.48 Ga); (D) IOM of the Orgueil meteorite; (E) and (F) samples a and b, respectively, of theJosefsdal Chert, Barberton Greenstone Belt, South Africa. The dotted lines in A, B, D, E highlight the shape of the proton ridge in biogenicand meteoritic IOM. All 2D HYSCORE spectra were recorded at a s value of 136 ns, with 256 � 256 points in both t1 and t2 dimensions.


only the 31P signal. This result strongly suggests that the IOMsof these twoArchean black cherts have a biological origin andunderwent a similar metamorphic history. This biological ori-gin of the IOM is well documented for the Josefsdal chert(Westall et al., 2015). On the contrary, the anomalous IOMpresent in slice 25 (horizon F0) of sample a exhibits a peculiarHYSCORE spectrum (Fig. 7E), with a combination of char-acteristics of biological and meteoritic IOMs. In particular,it shows the linear 1H ridge and the extended and relativelyintense signal of 13C characteristic of chondritic IOM, whilethe double-frequency 31P signal at 12 MHz is characteristicof fossilised IOM (Gourier et al., 2013).

In summary, two independent spectroscopic characteris-tics derived from cw-EPR and pulse-EPR measurementssupport the hypothesis that IOM of horizon F in samplea is a mixture of extraterrestrial hydrogenated organic mat-ter and of biogenic dehydrogenated organic matter, the lat-ter being present in small amounts throughout the sample.

4.4. Origin of the EPR signal at g = 3

In addition to the IOM EPR line presentingcharacteristics of extraterrestrial organic matter, horizonF (slices 23–26 of sample a) also features an EPR line at

g = 3 (Fig. 3C). This signal does not appear in other slicesof sample a, and was detected neither in the 27 fragments ofsample b, nor in any carbonaceous cherts of various agespreviously studied by EPR (Skrzypczak-Bonduelle et al.,2008). Examination of Fig. 4A (red symbols) shows thatthe intensity profile of the g = 3 signal through horizon Fexhibits two maxima clearly corresponding to lamina F0

and F00 0. The large deviation of the g-factor from the freeelectron value 2.0023 implies that the g = 3 signal cannotbe due to organic radicals and must therefore be due to amineral phase containing transition metals (Wertz andBolton, 1986). Furthermore, to the best of our knowledge,g-values between 4.3 (Fe3+ impurities in low symmetry) and�2.3 (Ni2+ and Cu2+) are rare in mineral compounds. Con-sequently, the presence of both extraterrestrial organic mat-ter and an anomalous EPR signal at g = 3 in the same thinsedimentary horizon F prompted us to explore a possibleextraterrestrial origin for this g = 3 signal. In favor of thishypothesis, spinel particles in the fusion crust of Allendemeteorite give a signal with a large g-value �2.5 (Griscomet al., 1999). Moreover Smart (1954) demonstrated thatNi-Fe-Al spinels can exhibit g-values covering the range1.5–5.8 depending on their metal composition. In order toconstrain the mineral compounds producing the signal at


g = 3, the chemical composition of horizon F and the sur-rounding sedimentary horizons were analyzed by PIXEspectroscopy.

4.4.1. PIXE analysis of horizon F

The elemental composition of horizon F in sample a wasmapped by the micro-PIXE method. The scanned area isrepresented by a white rectangular box in the optical imageof sample a in Fig. 4, which corresponds approximately toslices 20–30 of the EPR analyses. The image in Fig. 8A isa magnified view of the same surface area of sample a as out-lined in Fig. 4. A narrow chert vein (penecontemporaneousand possibly hydrothermal in origin) crossing lamina F0 andF00 0 is visible in the rectangular box. Representative PIXEspectra of the vein, of laminae F0, F00 0 and of a tuffaceoussediment are shown in Figs. C1–C4 of the SupplementaryInformation. Elemental maps for Fe, Ni and Cr qualita-tively show that the vein is enriched in Fe and Cr, while lam-inae F0 and F00 0 are enriched in Fe and Ni (Fig. 8A). In orderto better determine the composition of these three domains,the 25 PIXE spectra recorded for each of the 250 horizontalpositions (i.e. parallel to the stratification) were groupedtogether prior fitting in order to improve count statisticsand S/N ratio without sacrificing horizontal resolution.The obtained sensitivity was 200 lg/g for S, 12 lg/g forCr, 4 lg/g for Fe and 7 lg/g for Ni. The resulting 250-point elemental linescan across horizon F and the surround-ing tuffaceous sediment was further examined by selectingthree separated domains marked by vertical arrows inFig. 8B. Data from these three domains are represented inthe compositional plots in Fig. 9: lamina F00 0 (red squares),vein (green diamonds), lamina F0 (blue triangles), and thesurrounding tuffaceous layers referred to as the bulk (blackdots). The Ni/Cr plot (Fig. 9A) allows a clear distinction tobe made between the vein enriched in Cr and the F0 and F00 0

laminae, which contain almost the same Cr content as thebulk. In contrast, the two laminae are characterized byhigher Ni concentrations. While a previous Raman investi-gations reported the presence of small amounts of micro-crystalline (<5mm) pyrite in the Josefsdal Cherts (Westallet al., 2015), the Fe/S plot of sample a shows that data forlaminae F0 and F00 0 and the vein clearly lie outside thedomain occupied by iron sulphides (Fig. 9B). Furthermore,the Fe/Ni plot shows a marked distinction between thehydrothermal vein and the F0 and F00 0 laminae (Fig. 9C).Since pyrite may be excluded as the principal iron bearingmineral in lamina F0 and F00 0, it is more likely that Fe andNi are mainly present in spinels as such compounds mayexhibit large g-values (Smart, 1954).

3

Fig. 8. (A) Optical image of horizon F with laminae F0 and F00 0. The chebox represents the area scanned by the proton beam (5 mm � 0.5 mm) byshown on the bottom. (B) Projection on x-axis of data for Fe and Ni. Eagrouped over the 25 vertical pixels for each horizontal position of thedelineated by blue, red and green vertical lines, respectively. (For interprereferred to the web version of this article.)

4.4.2. The ‘‘cosmic” nanospinel model

Magnetic spinels are known to be formed in extraterres-trial objects during their entry in the Earth’s atmosphere,and are sometimes referred to as cosmic spinels (Robinet al., 1992). Such spinels from large impacts have beenidentified in spherules at the Cretaceous-Tertiary (K-T)boundary (Montanari et al., 1983), as well as in Archeanspherules beds from the BGB (Byerly and Lowe, 1994).They originate from liquid droplets formed at high temper-atures that quenched into glassy spherules (Gayraud et al.,1996). Among other characteristics, cosmic spinels differfrom terrestrial spinels by their higher Fe3+ and Ni, Cr con-tents (Smit and Kyte, 1984; Robin et al., 1992, 1993; Barnesand Roeder, 2001). However, apart from an EPR analysisof titanomagnetite nanoparticles in K-T spherules fromthe floor of the Chicxulub crater (Griscom et al., 1999), itseems that there have been no other EPR investigationsof spinels related to extraterrestrial objects.

The general composition of spinel-type ferrites is(M2+)A(Fe3+2 )BO4, where M represents a divalent metal.In normal spinels, divalent cations occupy the tetrahedralA-sites and trivalent cations the octahedral B-sites of thestructure. In inverse spinels, half of the B-sites are occupiedby the M2+ ions and the A-sites by Fe3+ ions, i.e.(Fe3+)A(M2+Fe3+)BO4. Spinel-type ferrites are generallyferrimagnetic compounds with the A-sites and B-sites con-stituting two antiferromagnetically coupled sublattices. Thenon-compensation of the magnetic moments of the A and B

sublattices gives a net magnetic moment M!

per unit cell

related to the total spin S!

of a unit cell by M!¼ �gb S

!,

where g is the g-factor and b the electron Bohr magneton.The magnetic moment and the total spin, hence the g-factor of spinel-type ferrites, are very sensitive to cationicdistributions between the A and B sites. Consequently theg-factor can be used to determine the cation site occupancyin the structure (Gorter, 1954; Smart, 1954). In a standardelectron magnetic resonance experiment, only the absolutevalue of the g-factor can be measured. The latter is givenby Smart (1954):

jgj ¼P

ixigiSi

� �A� P

ixigiSi

� �BP

ixiSi

� �A� P

ixiSi

� �B

��

��ð1Þ

where xi, gi, and Si are the fractional occupancies, theg-factor and the spin quantum number of ion i in eitherA or B site, respectively.

Since Fe3+ in spinels may be partially substituted byAl3+ or Cr3+, we considered a Ni-Cr-Al ferrite ofcomposition (Fe3+)A(Fe2+1�xNi2+x Cr3+t Al3+z Fe3+1�z�t)

BO4 with

rt vein crossing horizon F is highlighted by dashed lines. The blacksteps of 20 lm (250 � 25 pixels). PIXE images of Fe, Ni and Cr arech compositional result was obtained from the fitting of the spectrabeam. Data corresponding to lamina F0 and F00 0 and the vein aretation of the references to colour in this figure legend, the reader is

Table 1g-factor and electron spin S of transition metal elements in A and Bsites of spinel structure.

Fe3+ Fe2+ Ni2+ Cr3+

Site A B B B Bg-factor 2.0 2.0 2.1 2.3 2.0Spin S 5/2 5/2 2 1 3/2

Fig. 9. Scatter plots of Ni/Cr (A), S/Fe (B) and Fe/Ni (C) PIXEanalysis of sample a, which show the distinction between laminaeF0 (blue triangles) and F00 0 (red squares), the vein (green diamonds),and the surrounding tuffaceous layers (bulk, black points). (Forinterpretation of the references to colour in this figure legend, thereader is referred to the web version of this article.)


0 � x � 1 and 0 � z + t � 1. This composition can be for-mally written in terms of a mixture of single ion oxidesas: (1-x) FeO + x NiO + z/2 Al2O3 + t/2 Cr2O3 + (1-z/2-t/2) Fe2O3. To reduce the number of degrees of freedomand, considering that NiFe2O4 (trevorite) and Fe2+(Fe3+)2-O4 (magnetite) are inverse spinels (Fleet, 1981; Patangeet al., 2011), Fe2+ and Ni2+ were constrained to occupy

the B sites only. The g-values and spin number for eachion are given in Table 1.

For a fixed value of chromium content t, Eq. (1) with g

= ±3 yields a linear relationship between the Al content zand the Ni content x, and non-linear relationships betweenFe3+/Fetotal = ð2� z� tÞ=ð3� x� z� tÞ and the Ni con-tent x. Details of the calculation are given in Supplemen-tary Information D. From these relationships, a firstgraph can be drawn correlating the Fe3+/Fetotal ratio withthe NiO content for a range of Al content z and Cr contentt (blue dots, Fig. 10A). Another graph in Fig. 10B corre-lates the weight percent of Al2O3 with the weight percentof NiO for fixed values of chromium content t (full linesfor g = +3 and dashed lines for g = �3). These graphsdefine iron redox ranges and Al, Ni and Cr compositionranges of a spinel which give |g| = 3. In particular this g-value is possible only for Fe3+/Fetotal larger than 50%,depending on the NiO content (Fig. 10A). Although aEPR line with |g| = 3 can occur for a broad range ofAl2O3, NiO and Cr2O3 compositions (full lines and dashedlines in Fig. 10B), this g-value cannot occur in the compo-sition range below the line defined by [Al2O3] = �0.565[NiO] + 13 (dotted line in Fig. 10B).

The fact that the Fe3+/Fetotal must be larger than 50% togive spinels with |g| = 3 explains why such g-values have notbeen observed in terrestrial spinels, which exhibit low Fe3+

content. As an example, the green rectangles in Fig. 10 rep-resent the composition range of 3.24 Ga Archaean detritalspinels from eroded komatiitic volcanic rocks, measuredby electron microprobe (Krull-Davatzes et al., 2010). Itappears that the Fe3+/Fetotal ratio as well as the Al2O3

and Cr2O3 contents of these spinels are clearly outside theexistence domains of |g| = 3. On the contrary, cosmic spi-nels identified in a spherule bed in the K-T boundary(Robin et al., 1992, 1993), represented by the pink areasof Fig. 10, show too high Fe3+/Fetotal ratios and too lowNiO contents to give an EPR signal at |g| = 3. ConcerningArchaean, Krull-Davatzes et al. (2010) analyzed the com-position of cosmic spinels in the 3.24 Ga impact spherulebed S3, Fig Tree Group of the BGB. Their compositionrange is represented by the blue areas in Fig. 10. Interest-ingly, the large domain of experimental Fe3+/Fetotal versusNiO content (blue area in Fig. 10A) measured for theseArchaean cosmic spinels intersects the calculated curves(blue dots in Fig. 10A) for NiO content between 10 and20 wt% and Fe3+/Fetotal in the range 55–80 % (Fig. 10A).This compatibility of the |g| = 3 signal of horizon F withthe composition measured for Archaean cosmic spinels issupported by the correlation between Al2O3 content andNiO content for various Cr2O3 contents (Fig. 10B).Comparison of the |g| = 3 existence domain with actualcompositions of Archaean cosmic spinels shows that a part

Fig. 10. Chemical composition of a Ni-Cr-Al ferrite spinelcompatible with g = ±3. (A) The bundle of blue lines representthe Fe3+ to total Fe ratio vs NiO weight percent for varying valuesof chromium content t between 0 and 1 and varying values ofaluminum content z, calculated from Eqs. (B.5a), (B.5b), and (B.7)to (B.9) of Appendice B. (B) Al2O3 weight percent vs NiO weightpercent for varying chromium content t between 0 and 1 by steps of0.2. Full and dashed lines correspond to g = +3 (Eq. (B.5a)) andg = �3 (Eq. (B.5b)), respectively. The domain in which the |g| = 3lines exist is delimited by dotted lines. Blue areas represent valuesmeasured by electron microprobe analysis of cosmic spinels in3.24 Ga impact spherule beds S3 in the Barberton Greenstone Belt(BGB; Krull-Davatzes et al., 2010). Pink areas represent valuesmeasured for cosmic spinels of the K-T boundary (Robin et al.,1992, 1993). Green rectangles represent the average composition ofdetrital spinels derived from the erosion of 3.24 Ga komatiiticvolcanic rocks (Krull-Davatzes et al., 2010). (For interpretation ofthe references to colour in this figure legend, the reader is referredto the web version of this article.)


of the latter should exhibit a g factor g = �3 (intersection ofthe blue area with dashed lines for t = 0.6 and 0.8 inFig. 10B) for NiO contents of 10–20 wt% and for high Crcontent (the t values between 0.6 and 0.8 in Fig. 10B corre-spond to 22–29 weight% of Cr2O3, respectively). It shouldbe pointed out that these NiO contents (�10–20 %) of

cosmic spinels compatible with |g| = 3 in Fig. 10A and10B were derived independently, suggesting that our deduc-tions are coherent. However the Cr2O3 content calculatedfor a g = �3 line (22–29%) is lower than the effectiveCr2O3 content of 30–55% measured for the cosmic spinelsof the S3 spherule bed (Krull-Davatzes et al., 2010). Despitethis discrepancy, the experimental correlation curvebetween Fe3+/Fetotal and Cr2O3 established by theseauthors (Fig. 6 of Krull-Davatzes et al., 2010) shows thatcosmic spinels with our Fe3+/Fetotal values in the range55–80 % (the case of F horizon) would exhibit Cr2O3 con-tents in the range �15–30%, which fits very well with ourcalculated values of 22–29 % necessary to attribute the |g|= 3 EPR signal to Ni-Cr-Al-Fe spinels.

Cosmic spinels are formed by a high degree of melting ofthe object as it falls towards the Earth’s surface, withFe3+/Fetotal ratios mainly determined by the O2 fugacityof the atmosphere (Toppani and Libourel, 2003).Krull-Davatzes et al. (2010) observed that, although theBGB cosmic spinels show a large dispersion of Fe3+/Fetotaland NiO contents, there is a positive correlation betweenFe3+/Fetotal and NiO. This is naturally imposed by the factthat Ni2+ ions replace Fe2+ ions in the spinel structure.Hence, an increase in Fe3+/Fetotal implies a higher NiO con-tent. This correlation is clearly shown by the bundle of cal-culated curves characterizing a |g| = 3 line in the diagram ofFig. 10A (blue dots). Taken together, the curves in Fig. 10Aand the comparison with the experimental Fe redox and Nicontents found in BGB cosmic spinels (Fig. 7 in Krull-Davatzes et al., 2010) point to a low atmospheric oxygenfugacity of 10�5 < fO2 < 10�4 bar for obtaining the spinelcomposition deduced for horizon F of the Josefsdal Chert.These values are compatible with the oxygen-poorArchaean atmosphere. On the other hand, the Fe3+/Fetotaland NiO content of spinels from K-T spherules (pink areain Fig. 10A) correspond to a present day, oxygen-rich,atmosphere (Toppani and Libourel, 2003), and excludeEPR signals with g = 3.

Finally, it is important to note that the EPR line at |g| =3 is sufficiently narrow to be detectable in horizon F, whichindicates that the corresponding spinel are superparamag-netic, implying that the size of these particles is at most afew tens of nm (Griscom, 1984). For larger particles, theEPR signal would be broadened and distorted (ferromag-netic resonance), which would strongly limit their detectionat low concentrations (Griscom, 1984).

In summary the coexistence, in horizon F of the Josefs-dal Chert, of hydrogenated organic matter of chondriticorigin and cosmic spinel nanoparticles is a strong line ofevidence that this sedimentary layer contains a relativelyhigh concentration of materials of extraterrestrial origin.

5. DISCUSSION

This EPR analysis of a thin, carbonaceous horizon in3.33 Ga-old strata sediments from Facies D of the JosefsdalChert, Barberton Greenstone Belt, shows two independentspectroscopic features pointing to an extraterrestrial originof a fraction of the preserved organic and inorganic matter:


(i) The cw-EPR lineshape and linewidth of IOM in hori-zon F of Facies D are similar to those of carbona-ceous chondrites, but different from those of IOMpreserved in cherts from other Archean formations,including Facies C of the Josefsdal Chert (Fig. 6).This finding is confirmed by pulse-EPR (HYSCORE)analysis of the IOM of horizon F, which exhibits 1Hand 13C signals with characteristics similar to those ofIOM of carbonaceous chondrites (Gourier et al.,2013). The presence of a 31P signal, a biosignature(Gourier et al., 2013), is also present indicating thatthe IOM in horizon F is a mixture of two compo-nents: biogenic IOM (hereafter referred to as B-IOM) similar to that present in other parts of theJosefsdal Chert (e.g. Westall et al., 2015), andextraterrestrial IOM (hereafter referred to as E-IOM) similar to that present in carbonaceouschondrites.

(ii) An anomalous cw-EPR signal at g = 3 in horizon Fis attributed to Ni-Cr-Al ferrite nanoparticles, co-located with the E-IOM. Their chemical compositionis compatible with cosmic spinels produced byextraterrestrial materials entering the ArchaeanEarth’s atmosphere.

It is difficult to envisage a single impact event preservingboth organic matter and spinel particles in such a thin sed-imentary layer. On the one hand, hydrogenated organicmatter can survive only if the temperature of the fallingmatter does not exceed a few hundred degrees. On the otherhand, cosmic spinels are formed by a high degree of meltingof the object, as it falls towards the Earth’s surface. In suchconditions, the organic matter would be invariablydestroyed. It must be stressed that the evidence of impactevents during the Archaean is mainly deduced from thepresence of spherule beds. On this basis, at least 8 majorimpactor events (>10 km diameter) are documented in sed-iments of the BGB between 3.47 and 3.24 Ga, occurring ata frequency of about one impactor every 30 Ma (Loweet al., 2014; Lowe and Byerly, 2018). One of these impactors(S6) is located in the slightly younger M3c chert, strati-graphically just above the Josefsdal chert (the stratigraphicequivalent of K3c in Lowe et al., 2014), and apparentlycaused significant destruction due to tsunamis. The brecciabeds below and above Facies D may have had a related ori-gin (Westall et al., 2015). While large impactors show cleargeological signatures on the Early-Mid Archaean record(tsunamis, spherule beds), smaller impactors were muchmore abundant, but would have produced more discretesignatures, such as slumping and liquefaction, eitherdirectly from impact shock and/or through associated seis-mic activity. Today, the largest flux of extraterrestrial mate-rials falling on Earth is represented by micrometeorites(Prasad et al., 2013; Love and Brownlee, 1991; Dartoiset al., 2018 and references therein). Similarly, the flux ofextraterrestrial materials during the Archaean, much largerthan today, was also certainly largely dominated bymicrometeorites (cf Maurette and Brack, 2006). However,such tiny particles leave no evident geological trace of theirpresence. Importantly, the temperature reached by

submillimeter particles during atmospheric entry stronglydepends on their size. Love and Brownlee (1991) calculatedthat temperatures can reach 1500 �C for the largest particlesizes, allowing melting and spinel formation (Toppani andLibourel, 2003), while particles with sizes smaller than10 lm experience sufficiently low temperature to alloworganic matter to survive. Consequently it appears likelythat a continuous flow of micrometeorites falling on theArchean Earth provided an enormous quantity of extrater-restrial materials, among which cosmic spinels and E-IOMrepresent two extreme poles. It would therefore not be sur-prising to find both cosmic spinels and E-IOM in Archaeansediments, intermixed with B-IOM and volcanic materials.Thus, it is strongly suspected that such E-IOM and cosmicspinels could be present in the whole of all the JosefsdalChert formation, however in amounts too small to bedetected by EPR or other analytical techniques becausediluted by the other components. In our opinion, what isexceptional is not the finding of extraterrestrial materialin Archaean sediments, but the fact that, during a veryshort period of time (corresponding to the deposition ofF horizon), the proportion of extraterrestrial materialswas sufficiently high to be detectable.

We thus propose that a cloud of particles, mixing vol-canic ash and extraterrestrial materials, was present in theambient atmosphere above the Josefsdal volcanic island.A specific concatenation of circumstances, detailed below,then led to a brief concentration of the extraterrestrialmaterials in one particular sedimentary layer. This scenar-ios does not exclude the likelihood that extraterrestrialmaterials containing both E-IOM and spinel particles inthese sediments, may also result from the reworking andweathering of previous micrometeorite falls. What couldthe first-order control on the local concentration of E-IOM and cosmic spinels in a near-shore volcano sedimen-tary formation? Comparison with modern ashfall deposits(Orton, 1996; Mastin et al., 2009; Kratzmann et al., 2010)suggests that an �10 cm thick bundle of tephra (represent-ing one eruptive phase), resulting from multiple eruptivepulses (events), is probably deposited in several hours todays, and the whole of the 0.6–1.2 m thick unit of green-grey and black laminated sediments (Unit 3/Facies D,Fig. 1) could have been deposited on a scale of weeks tomonths. The thin F horizon deposition that concerns us,which can be traced stratigraphically for several kilometers,was deposited in a very short time and immediately buriedby the following thin ash layer and thus sealed withoutbeing reworked by currents or waves. If our scenario ofwindblown debris from micrometeorite falls is correct,carbonaceous layers slightly above and below the sectionanalyzed could possibly be enriched to various degrees inE-IOM.

The scenario that we propose is the following. Facies Dwas formed primarily by passive settling of volcanic ashduring explosive activity, the ash clouds comprising mostlyvolcanic material as well as extraterrestrial materials fromcontinuous flux of micrometeorites. During passive settlingin the water, the heaviest particles sediment first, followedthe lightest ones, resulting in the vertical graded textureobserved in facies D. Thus the finest part of the ash cloud,


including the extraterrestrial components, would settle outlast. As a matter of fact, the presence of very fine particlesin the atmosphere induces the condensation of water dro-plets, forming a fine rain. Consequently it was only duringa very short and quiet period of remission between volcanicevents, combined with no current or wave activity, that thedust was gently deposited. In such a geological setting, thepresence of extraterrestrial dust would be detectable onlybecause these conditions were present simultaneously,resulting in a local concentration of extraterrestrial materialin horizon F. Under more agitated conditions (strong cur-rent and wave action and ashfall), the extraterrestrial dustwould have been dynamically intermixed with and dilutedby volcanic ash in the atmosphere and in the sediments,thus preventing its detection by EPR.

Importantly, the preservation of hydrogenated E-IOMin 3.33 Ga-old sediment indicates that the molecular struc-ture of the organic precursors was partially preserved formore than 3 billion years owing to the rapid silicificationand lithification of the sedimentary layers. This hypothesismay appear counterintuitive, as it is well documented thatthe H/C ratio of IOM decreases with increasing maturation(see Vandenbroucke and Largeau, 2007 for a review). Asboth biogenic and meteoritic organic matter trapped inthe Archean sediment underwent the same thermodynamicand kinetic conditions (temperature, pressure, time), theyshould be at the same stage of maturation and, in particu-lar, should exhibit the same amount of hydrogen loss. Thefact that B-IOM has lost its hydrogen (facies C, for exam-ple) while E-IOM (horizon F of facies D) conserved a sig-nificant hydrogen fraction indicates that differences inmolecular structure of the organic precursor may be animportant factor controlling its geological evolution(Vandenbroucke and Largeau, 2007). We suggest that thehighly branched character of aliphatic chains in chondriticIOM (Gardinier et al., 2000; Derenne and Robert, 2010)could have hindered or slowed down their aromatization.Thus, for steric reasons, some C-H bonds could have beenpreserved in the E-IOM, while the more linear (lessbranched) organic chains in precursors of biological origin(mainly lipids) implies that their aromatization (and there-fore hydrogen loss) may occur more readily for the samesteric reasons. This hypothesis needs to be tested by labora-tory experiments (accelerated maturation of E-IOM and ofbiogenic B-IOM) but suggests the exciting possibility thatthe memory of the origin of the organic precursor couldbe identified by EPR analysis of IOM even after several bil-lion years.

Our documentation of the physical presence of E-IOMin early terrestrial sediments is significant in that it underli-nes the importance of the flux of extraterrestrial carbon tothe early Earth, already modelled to be high, �5 � 1024 gaccreted in the 300 My after the Moon-forming impact(e.g. Maurette and Brack, 2006). The high content of car-bon molecules in micrometeorites, comprising both solubleorganic matter (ESOM), such as PAHs and amino acids, aswell as IOM, would have supplied the early Earth with amajor fraction of organic material necessary for seedingthe emergence of life (Brack et al., 2011; Dartois et al.,2018). As we have seen above, this extraterrestrial matter

is so fine that it is readily disseminated in the seawaterand in the sediments deposited in the early oceans. Incorpo-rated in the sediments, the ESOM will have been rapidlydissolved while the more refractory E-IOM fraction wouldhave formed part of the sediments. Recycling of the sedi-ments containing E-IOM by tectonic activity, combinedwith hydrothermal leaching within the upper crust, wouldhave resulted in cracking of the molecules to form smallerentities, expelled by hydrothermal fluids into the seawaterand sediments adjacent to hydrothermal vents, that couldparticipate in the prebiotic processes leading to the emer-gence of protolife or primitive cells (Westall et al., 2018).

6. CONCLUSIONS

We have shown that a 1 mm-thick lamina of carbona-ceous matter in rhythmically-graded tuffaceous sedimentsdeposited in a littoral (shoreface to upper shoreface) envi-ronment exhibits EPR characteristics of IOM signature(cw-EPR spectrum lineshape and linewidth; pulse-EPR(HYSCORE) spectrum of 1H and 13C) that are similar tothose observed in the IOM of carbonaceous chondrites.These extraterrestrial IOM signatures are co-located withthe EPR signature of submicrometric-sized particles ofNi-Cr-Al ferrite spinels, which can be formed during entryof extraterrestrial objects into an oxygen-poor atmosphere.We conclude that there is strong evidence for extraterres-trial influence in 3.33 Ga-old volcanic sediments from theJosefsdal Chert. The IOM-rich laminae represent the topsof upward-fining sequences of airborne volcanic ash. Opti-cal microscopy and EPR study of all carbon-rich laminae inthe sample analyzed suggests that, apart from the particularlamina (F0) in question, the carbonaceous matter is of ter-restrial biological origin. In lamina F0 the IOM is signifi-cantly diluted by extraterrestrial carbon.

The detection of extraterrestrial carbonaceous matter inEarly Archean sediments underlines the importance of theflux of extraterrestrial carbon to the early Earth from theHadean onwards, as soon as the surface of the planetbecame habitable, and its contribution to the prebiotic pro-cesses that led to the emergence of life. The co-existence oftwo phases of carbon, one extraterrestrial and one biogenic,in the same sedimentary deposit emphasizes the challengesfacing in situ study of carbonaceous materials on otherplanetary bodies, such as Mars, where the search for tracesof life is primarily based on the characteristics of theorganic matter and where an extraterrestrial componenthas already been revealed (Eigenbrode et al., 2018).

ACKNOWLEDGMENTS

We express our gratitude to Cecile Engrand and the other tworeviewers for their illuminating comments and reviews which havegreatly improved the text, calling our attention on many detailsthat needed to be clarified. We also acknowledge funding fromthe French Space Agency (CNES) and from the European Com-munity’s Seventh Framework Programme (FP7/2007-2013) underGrant Agreement no 607297 (MASE project). We thank C.Pacheco for giving access to the Accelerateur Grand Louvre pourl’Analyse Elementaire (AGLAE) facility and Q. Lemasson, L.Pichon and B. Moignard for their help with the PIXE measure-


ments. Support from the EPR research network RENARD (FR-CNRS 3443) for providing access to the pulse EPR equipment isgratefully acknowledged. We thank Yann Le Du for his help inthe sample preparation.

APPENDIX A. SUPPLEMENTARY DATA

Supplementary data to this article can be found online athttps://doi.org/10.1016/j.gca.2019.05.009.

REFERENCES

Barnes S. J. and Roeder P. L. (2001) The range of spinelcompositions in terrestrial mafic and ultramafic rocks. J. Petrol.42, 2279–2302.

Binet L., Gourier D., Derenne S. and Robert F. (2002) Hetero-geneous distribution of paramagnetic radicals in insolubleorganic matter from the Orgueil and Murchison meteorites.Geochim. Cosmochim. Acta 66, 4177–4186.

Binet L., Gourier D., Derenne S., Robert F. and Ciofini I. (2004)Occurrence of abundant diradicaloid loieties in the insolubleorganic matter from the Orgueil and Murchison meteorites: afingerprint of its extraterrestrial origin ? Geochim. Cosmochim.

Acta 68, 881–891.Boehnke P. and Harrison T. M. (2016) Illusory late heavy

bombardments. Proc. Natl. Acad. Sci. U.S.A. 2; 113(39),10802–10806.

Botis S. M. and Pan Y. (2009) Theoretical calculations of[AlO4/M

+]0 defects in quartz and crystal-chemical controls onthe uptake of Al. Min. Mag. 73, 537–550.

Bottke W. F., Vokrouhlicky D., Minton D., Nesvorny D.,Morbidelli A., Brasser R., Simonson B. and Levison H. F.(2012) An Archean heavy bombardment from a destabilizedextension of the asteroid belt. Nature 485, 78–81.

Bourbin M., Gourier D., Derenne S., Binet L., Le Du Y., WestallF., Kremer B. and Gautret P. (2013) Dating carbonaceousmatter in Archean cherts by Electron Paramagnetic Resonance.Astrobiology 13, 151–162.

Brack A. (2011) Extraterrestrial delivery of organic compounds.Springer, pp. 821–828.

Byerly G. R. and Lowe D. R. (1994) Spinels from Archean impactspherules. Geochim. Cosmochim. Acta 58, 3469–3486.

Calligaro T., Coquinot Y., Pichon L. and Moignard B. (2011)Advances in elemental imaging of rocks using the AGLAEexternal microbeam. Nucl Instr. and Meth. B 269, 2364–2372.

Campbell J. L., Boyd N. I., Grassi N., Bonnicka P. and Maxwell J.A. (2010) The Guelph PIXE software package IV. Nucl Instr.

and Meth. B 268, 3356–3363.Cooper G. W., Kimmich N. and Belizle W., et al. (2001)

Carbonaceous meteorites as a source of sugar-related organiccompounds for the early Earth. Nature 414, 879–883.

Cosmidis J., Benzerara K., Guyot F., Skouri-Panet F., Duprat E.,Ferard C., Guigner J. M., Babonneau F. and Coelho C. (2015)Calcium-phosphate biomineralization induced by alkalinephosphataseactivity in Escherichia coli: Localization, kinetics,and potential signatures in the fossil records. Front. Earth Sci.

3, article 84.Dartois E., Engrand C., Duprat J., Godard M., Charon E.,

Delauche L., Sandt C. and Borondics F. (2018) Dome Cultracarbonaceous Antarctic micrometeorites Infrared andRaman fingerprints. A & A 609, A65.

Delpoux O., Gourier D., Vezin H., Binet L., Derenne S. andRobert F. (2011) Biradical character of the D-rich carriers in

the insoluble organic matter of carbonaceous chondrites : a relicof the protoplanetary chemistry. Geochim. Cosmochim. Acta.

75, 326–336.Derenne S. and Robert F. (2010) Model of molecular structure of

the insoluble organic matter isolated from Murchison mete-orite. Meteoritic Planet. Sci. 45, 1461–1475.

Dickneider T. A., Scull S., Whelan J. K. and Blough N. V. (1997)EPR study of kerogens from Middle Valley, Northern Juan deFuca Ridge. Org. Geochem. 26, 341–352.

Duchesne J., Depireux J. and Litt C. (1964) Concerning the natureof free radicals in the Cold Bokkeveld meteorite. Geoch. Int. 1,1022–1024.

Duprat J., Dobrica E., Engrand C., Aleon J., Marrocchi Y.,Mostefaoui S., Meibom A., Leroux H., Rouzaud J. N.,Gounelle M. and Robert F. (2010) Extreme deuterium excessesin ultracarbonaceous micrometeorites from central Antarcticsnow. Science 328, 742–745.

Eigenbrode J. L., Summons R. E., Steele A., Freissinet C., MillanM., Navarro-Gonzalez R., Sutter B., McAdam A. C., Franz H.B., Glavin D. P., Archer, Jr., P. D., Mahaffy P. R., Conrad P.,Hurowitz J. A., Grotzinger J. P., Gupta S., Ming D. W.,Sumner D. Y., Szopa C., Malespin C., Buch A. and Coll P.(2018) Organic matter preserved in 3-billion-year-old mud-stones at Gale crater, Mars. Science 360, 1096–1101.

Fleet M. E. (1981) The structure of magnetite. Acta Cryst. B37,917–920.

Foucher F., Guimbretiere G., Bost N. and Westall F. (2017)Petrographical and mineralogical applications of raman mapping

(Chapter 8). InThechOpen Limited, London, pp. 163–180.Gardinier A., Derenne S., Robert F., Behar F., Largeau C. and

Maquet J. (2000) Solid state CP/MAS 13C NMR of theinsoluble organic matter of the Orgueil and Murchison mete-orites: quantitative study. Earth Planet. Sci. Lett. 184, 9–21.

Gayraud J., Robin E., Rocchia R. and Froget L. (1996) Forma-tion condition of oxidized Ni-rich spinel and their relevance tothe K/T boundary event. Geol. Soc. Am. Spec. Pap. 307, 425–443.

Genda H., Brasser R. and Mojzsis S. J. (2017) The terrestrial lateveneer from core disruption of a lunar -sized impactor. EarthPlanet. Sci. Lett. 480, 25–32.

Glikson A. Y., Allen C. and Vickers J. (2004) Multiple 3.47-Ga-oldasteroid impact fallout units, Pilbara Craton, Western Aus-tralia. Earth Planet. Sci. Lett. 221, 383–396.

Gomes R., Levison H. F., Tsiganis K. and Morbidelli A. (2005)Origin of the cataclysmic Late Heavy Bombardment period ofthe terrestrial planets. Nature 435, 466–469.

Gorter E. W. (1954) Ionic distribution deduced from the g-factor ofa ferrimagnetc spinel: Ti4+ in fourfold coordination. Nature

173, 123–124.Gourier D., Robert F., Delpoux O., Binet L., Vezin H., Moissette

A. and Derenne S. (2008) Extreme deuterium enrichment oforganic radicals in the Orgueil meteorite: revisiting the inter-stellar interpretation? Geochim. Cosmochim. Acta 72, 1914–1923.

Gourier D., DelpouxO., Skrzypczak-Bonduelle A., Binet L., CiofiniI. and Vezin H. (2010) EPR, ENDOR and HYSCORE study ofthe structure and the stability of vanadyl-porphyrin complexesencapsulated in silica: Potential paramagnetic biomarkers for theorigin of life. J. Phys. Chem. B 114, 3714–3725.

Gourier D., Delpoux O., Binet L. and Vezin H. (2013) Nuclearmagnetic biosignatures in the carbonaceous matter of ancientcherts: comparison with carbonaceous meteorites. Astrobiology13, 932–947.

Govindaraju K. (1982) Report (1967–1981) on four ANRT rockreference samples: diorite DR-N, serpentine UB-N, bauxite BX-N and disthene DT-N. Geostandards Newsletter 6(1), 91–159.


http://refhub.elsevier.com/S0016-7037(19)30266-2/h0005




























































































































Griscom D. L., Beltran-Lopez V., Merzbacher C. I. and Bolden E.(1999) Electron spin resonance of 65-million-year-old glassesand rocks from the Cretaceous-Tertiary boundary. J. Non-

Cryst.Solids 253, 1–22.Griscom D. L. (1980) Ferromagnetic resonance of fine grained

precipitates in glass: a thumbnail review. J. Non-Crystalline

Solids 42, 287–296.Griscom D. L. (1984) Ferromagnetic resonance of precipitated

phases in natural glasses. J. Non-Crystalline Solids 67, 81–118.Kratzmann D. J., Carey S. N., Fero J., Scasso R. A. and Naranjo J.

A. (2010) Simulations of tephra dispersal from the 1991explosive eruptions of Hudson volcano, Chile. J. Volcanol.

Geoth. Res. 190, 337–352.Krull-Davatzes A. E., Byerly G. R. and Lowe D. R. (2010)

Evidence of a lowe-O2 Archean atmosphere from nickel-richchrome spinels in 3.24 impact spherules, Barberton greenstonebelt South Africa. Earth Planet. Sci. Lett. 296, 319–328.

Li J., Benzerara K., Bernard S. and Beyssac O. (2013) The linkbetween biomineralization and fossilization of bacteria: insightsfrom field and experimental studies. Chem. Geol. 359, 49–69.

Love S. G. and Brownlee D. E. (1991) Heating and thermaltransformation of micrometeoroids entering the Earth’s atmo-sphere. Icarus 89, 26–43.

Lowe D.R., Byerly G.R. and Heubeck C. Geologic map of thewest-central Barberton Greenstone Belt: Geological Society ofAmerica Map and Chart Series MCH103, 1 sheet, scale 1:25000, 2012.

Lowe D. R., Byerly G. R. and Kyte F. T. (2014) Recentlydiscovered 3.42-3.23 Ga impact layers, Barberton Belt, SouthAfrica: 3.8 Ga detrital zircons, Archean impact history, andtectonic implications. Geology 42, 747–750.

Lowe D. R. and Byerly G. R. (2018) The terrestrial record of LateHeavy Bombardment. New Astron. Rev. 81, 39–61.

Mastin I. G., Guffanti M., Servranckx R., Webley P., Barsotti S.,Dean K., Durant A., Ewert J. W., Neri A., Rose W. I.,Schneider D., Siebert I., Stundert B., Swanson G., Tupper A.,Volentik A. and Waythomas C. F. (2009) A multidisciplinaryeffort to assign realistic source parameters to models of volcanicahs-cloud transport and dispersion during eruptions. J. Vol-canol. Geoth. Res. 186, 10–21.

Marty B., Alexander C. M. O. D. and Raymond S. N. (2013)Primordial origins of Earth’s carbon. Rev. Mineral. Geochem.

75, 149–181.Maurette M. and Brack A. (2006) Cometary petroleum in Hadean

time? Meteorit. Planet. Sci. 41, 5247.Montanari A., Hay R. L., Alvarez W., Asaro F., Michel H. V.,

Alvarez L. W. and Smit J. (1983) Spheroids at the Cretaceous-Tertiary boundary are altered impact droplets of basalticcomposition. Geology 11, 668–671.

Morbidelli A., Lunine J. I., O’Brien D. P., Raymond S. N. andWalsh K. J. (2012) Building terrestrial planets Annu. Rev. EarthPlanet. Sci. 40, 251–275.

Mrozowski S. (1988a) ESR study of carbonization and coalificationprocesses part I: carbonaceous compounds.Carbon 26, 521–529.

Mrozowski S. (1988b) ESR study of carbonization and coalifica-tion processes part II: biological materials. Carbon 26, 531–541.

Orton G. J. (1996) Volcanic environments. In Sedimentary Envi-

ronments: Processes, Facies and Stratigraphy (ed. H. G.Reading). Wiley, p. 704.

Pasek M. and Lauretta D. (2008) Extraterrestrial flux of potentialprebiotic C, N and P to the early Earth. Orig. Life Evol. Biosph.

38, 5–21.Patange S. M., Shirsath S. E., Jangam G. S., Lohar K. S., Jadhav

S. S. and Jadhav K. M. (2011) Rietveld structure refinement,cation distribution and magnetic properties of Al3+ substitutedNiFe2O4 nanoparticles. J. Appl. Phys. 109, 053909.

Pichon L., Moignard B., Lemasson Q., Pacheco C. and Walter P.(2014) Development of a multi-detector and a systematicimaging system on the AGLAE external beam. Nucl Instr.

and Meth. B 318, 27–31.Pinti D. L. (2005) The formation and evolution of the oceans. In

Lectures in Astrobiology (eds. M. Gargaud, B. Barbier, H.Martin and J. Reisse). Springer, Berlin.

Prasad M. S., Rudraswami N. G. and Panda D. K. (2013)Micrometeorite flux on Earth during the last �50,000 years. J.Geophys. Res. Planets 118, 2381–2399. https://doi.org/10.1002/2013JE004460.

Remusat, L., 2014. Organic material in meteorites and the link tothe origin of life. BIO web of conferences, 2, 03003. http://doi.org/10.1051/bioconf/20140203001.

Retcofsky H. L., Stark J. M. and Friedel R. A. (1968) Electron spinresonance in American coals. Anal. Chem. 40, 1699–1704.

Robin E., Bonte Ph., Froget L., Jehanno C. and Rocchia R. (1992)Formation of spinels in cosmic objects during atmosphericentry: a clue to the Cretaceous-Tertiary boundary event. EarthPlanet. Sci. Lett. 108, 181–190.

Robin E., Froget L., JehannoC. andRocchlaR. (1993) Evidence fora K/T impact event in the Pacific Ocean. Nature 363,615–617.

Ryan C. G. (1995) The Nuclear Microprobe as a probe of earthstructure and geological processes. Nucl. Instr. and Meth. B 104,377–394.

Schmitt-Kopplin P., Gabelica Z., Gougeon R. D., Fekete A.,Kanawati B., Harir M., Gebefuegi I., Eckel G. and HertkornN. (2010) High molecular diversity of extraterrestrial organicmatter in Murchison meteorite revealed 40 years after its fall.PNAS 107, 2763–2768.

Schoenberg R., Kamber B. S., Collerson K. D. and Moorbath S.(2002) Tungsten isotope evidence from � 3.8-Gyr metamor-phosed sediments for early meteorite bombardment of theearth. Nature 418, 403–405.

Schultz K. F. and Elofson R. M. (1965) Electron spin resonance oforganic matter in the Orgueil meteorite. Geochim. Cosmochim.

Acta. 29, 157–160.Schweiger A. and Jeschke G. (2001) Principle of Pulsed Electron

Paramagnetic Resonance. Oxford University Press, Oxford.Sephton M. A. (2002) Organic compounds in carbonaceous

meteorites. Natl. Prod. Rep. 19, 292–311.Simonson B. M. and Glass B. P. (2004) Spherule layers – records of

ancient impacts. Ann. Rev. Earth Planet. Sci. 32, 329–361.Skrzypczak-Bonduelle A., Binet L., Delpoux O., Vezin H., Derenne

S., Robert F. and Gourier D. (2008) EPR of radicals in primitiveorganic matter: a tool for the search of biosignatures of the mostancient traces of life. Appl. Magn. Reson. 33, 371–397.

Smart J. S. (1954) Cation distribution in mixed ferrites. Phys. Rev.94, 847–850.

Smit J. and Kyte F. T. (1984) Siderophile-rich magnetic spheroidsfrom the Cretaceous-Tertiary boundary in Umbria, Italy.Nature 310, 403–405.

Strasdeit H. (2005) New studies on the Murchison meteorite shedlight on the pre-RNA world. ChemBioChem 6, 801–813.

Tartese R., Chaussidon M., Gurenko A., Delarue F. and Robert F.(2017) Warm Archean oceans reconstructed from oxygenisotope composition of early-life remnants. Geochem. Persp.

Let. 3, 55–65. https://doi.org/10.7185/55 geochemlet.1706.Toppani A. and Libourel G. (2003) Factors controlling composi-

tions of cosmic spinels: Application to atmospheric entryconditions of meteoritic materials. Geochim. Cosmochim. Acta

67, 4621–4638.Uebersfeld J., Etienne A. and Combrisson J. (1954) Paramagnetic

resonance, a new property of coal-like materials. Nature 174,614.









































































https://doi.org/10.1002/2013JE004460

https://doi.org/10.1002/2013JE004460

http://doi.org/10.1051/bioconf/20140203001

http://doi.org/10.1051/bioconf/20140203001











































https://doi.org/10.7185/55geochemlet.1706









Uebersfeld J. and Erb E. (1956) A new effect of oxygen on theparamagnetic resonance of charcoal. Comptes Rendus Acad Sci.

(Paris) 243, 363–364.Vandenbroucke M. and Largeau C. (2007) Kerogen origin,

evolution and structure. Org. Geochem. 38, 719–833.Vinogradov A. P., Vdovykin G. P. and Marov I. N. (1964) Free

radicals in the Miggei meteorite. Geoch. Int. 1, 395–398.Wertz J. E. and Bolton J. R. (1986) Electron Spin Resonance:

Elementary Theory and Practical Applications. Chapman andHall, New York.

Westall F., de Ronde C. E. J., Southam G., Grassineau N., ColasM., Cockell C. and Lammer H. (2006) Implications of a 3.472-3.333 Gyr-old subaerial microbial mat from the Barbertongreenstone belt, South Africa for the UV environmentalconditions on the early Earth. Phil. Trans. R. Soc. B 361,1857–1875.

Westall F., Cavalazzi B., Lemelle L., Marrocchi Y., Rouzaud J. N.,Simionovici A., Salome M., Mostefaoui S., Andreazza C.,Foucher F., Toporski J., Jauss A., Thiel V., Southam G.,MacLean L., Wirick S., Hofmann A., Meibom A., Robert F.and Defarge C. (2011) Implications of in situ calcification forphotosynthesis in a �3.3 Ga-old microbial biofilm from theBarberton Greenstone Belt, South Africa. Earth Planet. Sci.

Lett. 310, 468–479.

Westall F., Campbell K. A., Breheret J. G., Foucher F., Gautret P.,Hubert A., Sorieul S., Grassineau N. and Guido D. M. (2015)Archean (3.33 Ga) microbe-sediment systems were diverse andflourished in a hydrothermal context. Geology 43, 615–618.

Westall F., Hickman-Lewis K., Hinman N., Gautret P., CampbellK. A., Breheret J. G., Foucher F., Hubert A., Sorieul S., DassA. V., Kee T. P., Georgelin T. and Brack A. (2018) Ahydrothermal-sedimentary context for the origin of life. Astro-biology 18, 1–35.

Weeks R. A., Magruder, III, R. H. and Stesman A. (2008) Reviewof some experiments in the 50 year saga of the E’ center andsuggestions for future research. J. Non-Cryst. Solids 354, 208–216.

Weil J. A. (1984) A review of Electron Spin Spectroscopy and itsapplication to the study of paramagnetic defects in crystallinequartz. Phys. Chem. Miner. 10, 149–165.

Yada T., Nakamura T., Takaoka N., Noguchi T., Terada K.,Yano H., Nakazawa T. and Kojima H. (2004) The globalaccretion rate of extraterrestrial materials in the last glacialperiod estimated from the abundance of micrometeorites inAntarctic glacier ice. Earth Planets Space 56, 67–79.

Associate editor: Christian Koeberl
















































extraterrestrial organic matter preserved in 3.33â€¯ga

Documents